Dissertation > Excellent graduate degree dissertation topics show

The Stereo Matching and De-interlacing Algorithm Based on CUDA

Author: YanRui
Tutor: DuZuo
School: Zhejiang University
Course: Information and Communication Engineering
Keywords: Stereo matching De-interlacing Motion adaptive Motion compensation CUDA
CLC: TP391.41
Type: Master's thesis
Year: 2010
Downloads: 356
Quote: 4
Read: Download Dissertation

Abstract


Stereo vision is a hot research topic in the field of computer vision , the passive depth measurement method . Stereo matching is an important step in the three-dimensional vision , but also one of the difficulties . In this paper, the real-time stereo matching algorithm . Stereo matching algorithm , local area - based matching method computing simple and neat , suitable for hardware acceleration to achieve and be able to get a dense disparity map . The matching window reasonable choice of a direct impact on the efficiency and accuracy of the match , but using the appropriate stereo matching acceleration method can greatly reduce the complexity of the algorithm . In a typical video surveillance scene , was walking around and fast-moving vehicles , surveillance cameras interlaced so that the edge of the persons and vehicles fuzzy video image quality . In order to get a clear surveillance video , you need to turn the interlaced signal into a progressive scan signal . Mainstream better deinterlacing methods are two kinds of motion adaptive and motion compensation , these two methods are carried out the study . Race 4 of the same polarity motion detection method to extract the motion information , the image is divided into a still area , movement area and the mixed region ; the still area directly Occasion and for the motion area with improved motion adaptive interpolation based on the edge , the mixed region based on the weighted average of the motion vector . Motion compensated de -interlacing algorithm using bi-directional motion estimation method , according to the amplitude of the motion vector field image block is divided into three categories, fast motion block using interpolation based on the edge of the Exchange , the slower linear average of the block along the motion vector , the intermediate the speed of the block is subdivided Analyzing the interpolation calculation using one of the above two methods . CUDA is NVIDIA Corporation released for GPU general computing development environment and software architecture , parallel processing can be realized by means of powerful computing capabilities of the GPU . This article is based on CUDA accelerated stereo matching and de-interlacing algorithm . In stereo matching to compare the different structure of the algorithm by the CUDA acceleration performance difference . De-interlacing algorithm algorithm targeted structural adjustment , motion-adaptive de-interlacing motion detection and motion compensated de-interlacing motion estimation CUDA acceleration , achieved good results .

Related Dissertations

  1. Study of Binocular Stereo Matching Algorithm and Algorithm Implementation Based on Multi-core.,TP391.41
  2. Spatial Temporal Refinement on Depth Video Processing,TP391.41
  3. Research of Finite Element Method on GPU,O241.82
  4. Research of Multimedia Enhancement Unit in the Embedded Processor,TP332
  5. CUDA-based regular expression matching system design and implementation,TP311.52
  6. White Light Interferometry for Fast Areal Surface Measurement Based on GPGPU,O439
  7. Research groups and evacuation simulation algorithm simulation system,TP391.9
  8. GPU-based book recommendation system Research and Implementation,TP391.3
  9. Scenario-based non- uniformity correction,TN215
  10. CUDA-based video fire detection system,TP391.41
  11. GPU-based reconstruction algorithm for X-rays accelerate research,TP391.41
  12. OpenCV-based digital image stabilization system to achieve,TP391.41
  13. GPU-based acceleration of neutral gas leak and rescue simulation study,TP391.41
  14. Research on Binocular Stereo Vision System for Indoor Mobile Robot on Complex Environment,TP242
  15. Real-Time Image Mosaic Based on CUDA,TP391.41
  16. Research of the Real-time Rock Deformation Monitoring System,TP274
  17. Stereo matching algorithm based on binocular stereo vision,TP391.41
  18. CPU GPU -based heterogeneous platforms string matching algorithm and implementation,TP301.6
  19. GPGPU and Image Matching Parallel Algorithm Based on SIFT,TP391.41
  20. High - performance computing systems based on particle simulation of the GPU,TP338
  21. Research of Acceleration Photon Mapping Based on GPU,TP391.41

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device
© 2012 www.DissertationTopic.Net  Mobile